diff --git a/bun.lock b/bun.lock
index a86571ba07..59bd1574e3 100644
--- a/bun.lock
+++ b/bun.lock
@@ -33,6 +33,7 @@
       "dependencies": {
         "@kobalte/core": "catalog:",
         "@opencode-ai/core": "workspace:*",
+        "@opencode-ai/schema": "workspace:*",
         "@opencode-ai/sdk": "workspace:*",
         "@opencode-ai/session-ui": "workspace:*",
         "@opencode-ai/ui": "workspace:*",
diff --git a/packages/app/package.json b/packages/app/package.json
index a4773e4da2..98f37c764f 100644
--- a/packages/app/package.json
+++ b/packages/app/package.json
@@ -46,6 +46,7 @@
   "dependencies": {
     "@kobalte/core": "catalog:",
     "@opencode-ai/core": "workspace:*",
+    "@opencode-ai/schema": "workspace:*",
     "@opencode-ai/sdk": "workspace:*",
     "@opencode-ai/session-ui": "workspace:*",
     "@opencode-ai/ui": "workspace:*",
diff --git a/packages/app/src/utils/id-wrap-boundary.test.ts b/packages/app/src/utils/id-wrap-boundary.test.ts
new file mode 100644
index 0000000000..5e26a1ab08
--- /dev/null
+++ b/packages/app/src/utils/id-wrap-boundary.test.ts
@@ -0,0 +1,58 @@
+import { describe, expect, test } from "bun:test"
+import { create } from "@opencode-ai/schema/identifier"
+
+const WRAP_BOUNDARY = 1786706395136
+const chars = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
+
+describe("app identifier seam: 48-bit wrap", () => {
+  test("ascending ids stay lexicographically ascending across the wrap boundary", () => {
+    const preWrap = create(false, WRAP_BOUNDARY - 1)
+    const postWrap = create(false, WRAP_BOUNDARY + 1)
+    expect(preWrap < postWrap).toBe(true)
+  })
+
+  test("descending ids stay lexicographically descending across the wrap boundary", () => {
+    const preWrap = create(true, WRAP_BOUNDARY - 1)
+    const postWrap = create(true, WRAP_BOUNDARY + 1)
+    expect(postWrap < preWrap).toBe(true)
+  })
+
+  test("same-millisecond ids are strictly ascending and unique", () => {
+    const ids = Array.from({ length: 10 }, (_, index) => create(false, WRAP_BOUNDARY + 1000 + index))
+    for (let i = 1; i < ids.length; i++) {
+      expect(ids[i - 1] < ids[i]).toBe(true)
+    }
+    expect(new Set(ids).size).toBe(ids.length)
+  })
+
+  test("same-millisecond descending ids are strictly descending and unique", () => {
+    const ids = Array.from({ length: 10 }, (_, index) => create(true, WRAP_BOUNDARY + 2000 + index))
+    for (let i = 1; i < ids.length; i++) {
+      expect(ids[i - 1] > ids[i]).toBe(true)
+    }
+    expect(new Set(ids).size).toBe(ids.length)
+  })
+
+  test("id format is 12 hex chars plus 14 base62 chars", () => {
+    const id = create(false, WRAP_BOUNDARY + 1)
+    expect(id).toHaveLength(26)
+    expect(id.slice(0, 12)).toMatch(/^[0-9a-f]{12}$/)
+    for (const char of id.slice(12)) {
+      expect(chars).toContain(char)
+    }
+  })
+
+  test("ids stay ascending when the clock regresses (latch absorbs regression)", () => {
+    const first = create(false, WRAP_BOUNDARY + 3000)
+    const regressed = create(false, WRAP_BOUNDARY + 3000 - 50)
+    const later = create(false, WRAP_BOUNDARY + 3000 + 50)
+    expect(first < regressed).toBe(true)
+    expect(regressed < later).toBe(true)
+  })
+
+  test("new-scheme ids sort below historical pre-wrap ids (comparisons must be time-based)", () => {
+    const now = create(false, Date.now())
+    const historical = create(false, WRAP_BOUNDARY - 1)
+    expect(now < historical).toBe(true)
+  })
+})
diff --git a/packages/app/src/utils/id.test.ts b/packages/app/src/utils/id.test.ts
new file mode 100644
index 0000000000..00d7b8af93
--- /dev/null
+++ b/packages/app/src/utils/id.test.ts
@@ -0,0 +1,75 @@
+import { describe, expect, test } from "bun:test"
+import { create } from "@opencode-ai/core/id/id"
+import { Identifier } from "./id"
+
+const prefixes = {
+  session: "ses",
+  message: "msg",
+  permission: "per",
+  user: "usr",
+  part: "prt",
+  pty: "pty",
+} as const
+
+function decodeTime(id: string): number {
+  const start = id.indexOf("_") + 1
+  return Number(BigInt("0x" + id.slice(start, start + 12)))
+}
+
+describe("Identifier", () => {
+  test("every prefix keeps the prefix_underscore_26-char shape", () => {
+    for (const prefix of ["session", "message", "permission", "user", "part", "pty"] as const) {
+      const ascending = Identifier.ascending(prefix)
+      const descending = Identifier.descending(prefix)
+      expect(ascending.startsWith(`${prefixes[prefix]}_`)).toBe(true)
+      expect(descending.startsWith(`${prefixes[prefix]}_`)).toBe(true)
+      for (const id of [ascending, descending]) {
+        expect(id).toHaveLength(prefixes[prefix].length + 1 + 26)
+        expect(id.slice(prefixes[prefix].length + 1, prefixes[prefix].length + 13)).toMatch(/^[0-9a-f]{12}$/)
+      }
+    }
+  })
+
+  test("ascending ids decode their time prefix back to raw wall-clock ms", () => {
+    // The legacy encoding shifted the millisecond value 12 bits and wrapped at
+    // 2026-08-14 (issue 271), so under it a live id no longer decodes near
+    // Date.now().
+    const before = Date.now()
+    const id = Identifier.ascending("message")
+    const after = Date.now()
+    const decoded = decodeTime(id)
+    expect(decoded).toBeGreaterThanOrEqual(before - 5000)
+    expect(decoded).toBeLessThanOrEqual(after + 5000)
+  })
+
+  test("app ids sort consistently with core ids for the same wall-clock", () => {
+    const appID = Identifier.ascending("message")
+    const coreID = create("msg", "ascending")
+    expect(Math.abs(decodeTime(appID) - decodeTime(coreID))).toBeLessThanOrEqual(5000)
+  })
+
+  test("same-millisecond bursts stay strictly ascending and unique", () => {
+    const ids = Array.from({ length: 20 }, () => Identifier.ascending("message"))
+    for (let index = 1; index < ids.length; index++) {
+      expect(ids[index - 1] < ids[index]).toBe(true)
+    }
+    expect(new Set(ids).size).toBe(ids.length)
+  })
+
+  test("descending ids stay strictly descending and unique", () => {
+    const ids = Array.from({ length: 20 }, () => Identifier.descending("session"))
+    for (let index = 1; index < ids.length; index++) {
+      expect(ids[index - 1] > ids[index]).toBe(true)
+    }
+    expect(new Set(ids).size).toBe(ids.length)
+  })
+
+  test("given id passes through when the prefix matches", () => {
+    expect(Identifier.ascending("session", "ses_abc123")).toBe("ses_abc123")
+    expect(Identifier.descending("message", "msg_abc123")).toBe("msg_abc123")
+  })
+
+  test("given id with a wrong prefix throws", () => {
+    expect(() => Identifier.ascending("message", "ses_abc123")).toThrow("ID ses_abc123 does not start with msg")
+  })
+})
diff --git a/packages/app/src/utils/id.ts b/packages/app/src/utils/id.ts
index dba7a8d951..0f4e363568 100644
--- a/packages/app/src/utils/id.ts
+++ b/packages/app/src/utils/id.ts
@@ -1,3 +1,5 @@
+import { create } from "@opencode-ai/schema/identifier"
+
 const prefixes = {
   session: "ses",
   message: "msg",
@@ -7,10 +9,6 @@ const prefixes = {
   pty: "pty",
 } as const
 
-const LENGTH = 26
-let lastTimestamp = 0
-let counter = 0
-
 type Prefix = keyof typeof prefixes
 export namespace Identifier {
   export function ascending(prefix: Prefix, given?: string) {
@@ -24,7 +22,7 @@ export namespace Identifier {
 
 function generateID(prefix: Prefix, descending: boolean, given?: string): string {
   if (!given) {
-    return create(prefix, descending)
+    return prefixes[prefix] + "_" + create(descending)
   }
 
   if (!given.startsWith(prefixes[prefix])) {
@@ -33,61 +31,3 @@ function generateID(prefix: Prefix, descending: boolean, given?: string): string
 
   return given
 }
-
-function create(prefix: Prefix, descending: boolean, timestamp?: number): string {
-  const currentTimestamp = timestamp ?? Date.now()
-
-  if (currentTimestamp !== lastTimestamp) {
-    lastTimestamp = currentTimestamp
-    counter = 0
-  }
-
-  counter += 1
-
-  let now = BigInt(currentTimestamp) * BigInt(0x1000) + BigInt(counter)
-
-  if (descending) {
-    now = ~now
-  }
-
-  const timeBytes = new Uint8Array(6)
-  for (let i = 0; i < 6; i += 1) {
-    timeBytes[i] = Number((now >> BigInt(40 - 8 * i)) & BigInt(0xff))
-  }
-
-  return prefixes[prefix] + "_" + bytesToHex(timeBytes) + randomBase62(LENGTH - 12)
-}
-
-function bytesToHex(bytes: Uint8Array): string {
-  let hex = ""
-  for (let i = 0; i < bytes.length; i += 1) {
-    hex += bytes[i].toString(16).padStart(2, "0")
-  }
-  return hex
-}
-
-function randomBase62(length: number): string {
-  const chars = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
-  const bytes = getRandomBytes(length)
-  let result = ""
-  for (let i = 0; i < length; i += 1) {
-    result += chars[bytes[i] % 62]
-  }
-  return result
-}
-
-function getRandomBytes(length: number): Uint8Array {
-  const bytes = new Uint8Array(length)
-  const cryptoObj = typeof globalThis !== "undefined" ? globalThis.crypto : undefined
-
-  if (cryptoObj && typeof cryptoObj.getRandomValues === "function") {
-    cryptoObj.getRandomValues(bytes)
-    return bytes
-  }
-
-  for (let i = 0; i < length; i += 1) {
-    bytes[i] = Math.floor(Math.random() * 256)
-  }
-
-  return bytes
-}